Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Amazon Rekognitionを用いてフォロワーの男女比を出す
Search
Sponsored
·
SiteGround - Reliable hosting with speed, security, and support you can count on.
→
Kazuki Ohashi
February 02, 2018
Technology
3k
1
Share
Amazon Rekognitionを用いてフォロワーの男女比を出す
Amazon RekogntionとFace APIを比べながら、フォロワーの男女比を出してみました。
Kazuki Ohashi
February 02, 2018
More Decks by Kazuki Ohashi
See All by Kazuki Ohashi
larave_vue_graphql_supplementation
kzkohashi
1
890
Introduction to using GraphQL for a bit
kzkohashi
0
240
フォロワーがどの雑誌に興味があるのか可視化してみる / magazine-score
kzkohashi
0
700
Laravelを始めて DDDを実践するまで
kzkohashi
2
2k
Other Decks in Technology
See All in Technology
[最強DB講義]推薦システム | 評価編
recsyslab
PRO
0
110
データ定義の混乱と戦う 〜 管理会計と財務会計 〜
wonohe
0
150
Percolatorを廃止し、マルチ検索サービスへ刷新した話 / Search Engineering Tech Talk 2026 Spring
visional_engineering_and_design
0
150
PicoRuby as a Multi-VM Operating System
kishima
1
210
「誰一人取り残されない」 AIエージェント時代のプロダクト設計思想 Product Management Summit 2026
mizushimac
1
1.6k
Good Enough Types: Heuristic Type Inference for Ruby
riseshia
1
310
AgentCore Managed Harness を使ってみよう
yakumo
2
230
AIを共同作業者にして書籍を執筆する方法 / How to Write a Book with AI as a Co-Creator
ama_ch
2
150
AIコーディング時代における、ソフトウェアサプライチェーン攻撃に対する防衛術(簡易版)
soysoysoyb
0
130
Choose your own adventure in agentic design patterns
glaforge
0
150
マルチプロダクトの信頼性を効率良く保っていくために
kworkdev
PRO
0
170
Expiration of Secure Boot Certificates for vSphere Virtual Machines
mirie_sd
0
110
Featured
See All Featured
JavaScript: Past, Present, and Future - NDC Porto 2020
reverentgeek
52
5.9k
Test your architecture with Archunit
thirion
1
2.2k
Building a Scalable Design System with Sketch
lauravandoore
463
34k
Beyond borders and beyond the search box: How to win the global "messy middle" with AI-driven SEO
davidcarrasco
3
110
From Legacy to Launchpad: Building Startup-Ready Communities
dugsong
0
200
Making the Leap to Tech Lead
cromwellryan
135
9.8k
New Earth Scene 8
popppiees
3
2.1k
30 Presentation Tips
portentint
PRO
1
280
The agentic SEO stack - context over prompts
schlessera
0
760
A Modern Web Designer's Workflow
chriscoyier
698
190k
Unlocking the hidden potential of vector embeddings in international SEO
frankvandijk
0
770
AI: The stuff that nobody shows you
jnunemaker
PRO
6
580
Transcript
"NB[PO3FLPHOJUJPOΛ༻͍ͯ ϑΥϩϫʔͷஉঁൺΛग़͢ WEBΤϯδχΞษڧձ #05 େڮ Ұथ @kzkohashi
େڮҰथ!L[LPIBTIJ ΠϯϑϧΤϯαʔϚʔέςΟϯά αʔόαΠυ ϑϩϯτத৺ झຯΧϨʔϥΠε
ϑΥϩϫʔͷใ ݟͨ͜ͱ͋Γ·͔͢ʁ
4/4ͷϑΥϩϫʔใ *OTUBHSBNΠϯαΠτ 5XJUUFS"OBMZUJDT
ଞਓͷΓ͍ͨʂ
Ͳ͏ΕͰ͖ΔͩΖ͏ʁ ϑΥϩϫʔ ϑΥϩϫʔ
إೝࣝΛͬͯஉঁྨਪʁ w จࣈϕʔεͰͷஉঁྨਪͦ͠͏ w 4/4ʹΑͬͯจࣈ͕গͳ͍߹͕͋Δ w ϓϩϑΟʔϧը૾͔Βஉ͔ঁ͔ͳΜͱ͘ͳ͘ Θ͔Γͦ͏
બఆख๏ͷௐࠪ ֎෦"1* ࣗલ 'BDF"1* $MPVE7JTPO 7JTVBM3FDPHOJUJPO 0QFO#3 ࣌ؒͱ͑Δ͓ۚΛߟ͑ͯ֎෦"1*͔Βબఆʂ "NB[PO3FLPHOJUJPO
͜Μͳײ͡ 'BDF"1* "NB[PO3FDPHOJUJPO
બఆൺΔ w ຕͷ4/4ͷϓϩϑը૾Λ༻ҙ w உੑຕ ঁੑຕ ͦͷଞຕ w 'BDF"1*ͱ"NB[PO3FLPHOJUJPOͰൺֱ ͋ΔํͷૉΒ͍͠ϒϩάͰࣝผ͕͔ͳΓΑ͔ͬͨʂ
ൺͨ݁Ռ ਖ਼ छྨ உੑ ঁੑ ͦͷଞ Amazon Rekognition 112(68%) 174(75%)
512(92%) Face API 75(46%) 121(52%) 522(94%) ୯७ͳਖ਼"NB[PO3FLPHOJUJPO͕ଟ͍
ൺͨ݁Ռʢޡʣ छྨ ঁੑ Amazon Rekognition 13 Face API 5 छྨ
உੑ Amazon Rekognition 5 Face API 3 ͕͑ঁੑͷը૾ ͕͑உੑͷը૾ ޡ'BDF"1*ͷ΄͏͕গͳͦ͏
'BDF"1* "NB[PO3FLPHOJUJPO ޱͳͲݟ͑ͳ͍ͱ'BDF"1*ݕग़ͮ͠Β͍ʁ ݕग़ࣦഊ ݕग़ޭ
'BDF"1* "NB[PO3FLPHOJUJPO 4/4ը૾ਅਖ਼໘Λ͋·Γ͍ͯͳ͍ͨΊݕग़ʹ͕ࠩग़ͨʁ ݕग़ࣦഊ ݕग़ޭ
'BDF"1* "NB[PO3FLPHOJUJPO ࣝผޭ͚ͨ͠Ͳ Ұਓ͚ͩݕग़ ೋਓݕग़͚ͨ͠Ͳ ࣝผࣦഊ ݕग़ͪ͠Ό͑'BDF"1*ڧ͍ɾɾɾ
'BDF"1* "NB[PO3FLPHOJUJPO ͜͏͍͏͜ͱ͋Δʢসʣ ࣝผޭ͚ͨ͠Ͳ ͲͪΒঁੑͱޡೝࣝ
બఆ·ͱΊ w 'BDF"1*ͷ΄͏͕ਫ਼ྑͦ͞͏ w ͨͩɺޱͱ͔ӅΕͨΓ͍ͯ͠Δ4/4ը૾ͷݕग़ѱ͗͢Δ w ࣝผͱݕग़ͷόϥϯεͰ"NB[PO3FLPHOJUJPOʹܾఆ ʢ୯७ͳਖ਼ͳΒΑ͔ͬͨʣ "NB[PO3FLPHOJUJPO
"NB[PO3FLPHOJUJPOͱ w "84͕ఏڙ͍ͯ͠Δɺը૾ࣝผαʔϏε w إೝ͚ࣝͩͰͳ͘ɺΦϒδΣΫτʢؠͱ͔ʣͷݕग़ ൺֱͱ͔Ͱ͖Δ w ͍͍ͳͱࢥͬͨͷɺஉঁࣝผʹ৴པ͕͍͍ͭͯΔ ͍·͞Β
(FOEFS\ 7BMVF.BMF $POpEFODF ^ ৴པ ࣗͨͪͷαʔϏεʹԠͯ͡ᮢܾΊΕΔ
ΞʔΩςΫνϟ ϓϩϑը૾ μϯϩʔυ ϑΥϩϫʔใ σϑΥϧτը૾͔ ൱͔ Ξοϓϩʔυ ˞ ˞͓͍ۚͬͨͳ͍ͷͰɺσϑΥϧτը૾ͷਓল͍ͯΔ
IUUQL[LPIBTIJIBUFOBCMPHDPNFOUSZ ʢQZUIPOΛͬͯ03#ͱ1FSDFQUVBM)BTIͰը૾ͷྨࣅΛൺͯΈΔ ৴པͷᮢ உঁใ "NB[PO3FLPHOJUJPO KTPO
݁Ռ ͑ ঁੑ உੑ ࠓճͷγεςϜ ˞ ঁੑ உੑ ˞αϯϓϦϯάϥϯμϜͰׂ̍ఔ
·ͱΊ w ը૾ࣝผ"1*৭ʑ͋Δ w ༻్ʹ߹Θͤͯࣝผͱݕग़ͷόϥϯεͰબ w ϓϩϑը૾͚ͩͰஉঁൺͱΕͨ
͓·͚ ͜ͷൺʹຊʹͳΔͷ͔ࢼͨ͠ छྨ உੑ ঁੑ ͦͷଞ ࣗͰ͑Δ 65%(97) 35%(52) 12
Amazon Rekognition 46%(36) 54%(42) 22 Face API 51%(21) 49%(20) 121 ͳΔ΄Ͳɾɾɾ